{"id":"https://openalex.org/W7225325521","doi":"https://doi.org/10.5281/zenodo.23974846","title":"A choice is not a lock: configuration-based tool disabling in an agent harness, evaluated with four language models","display_name":"A choice is not a lock: configuration-based tool disabling in an agent harness, evaluated with four language models","publication_year":2026,"publication_date":"2026-11-05","ids":{"openalex":"https://openalex.org/W7225325521","doi":"https://doi.org/10.5281/zenodo.23974846"},"language":"en","primary_location":{"id":"doi:10.5281/zenodo.23974846","doi":"https://doi.org/10.5281/zenodo.23974846","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.23974846","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.5281/zenodo.23974846","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5070693424","display_name":"Marcelo Cavicchioli","orcid":"https://orcid.org/0000-0002-5592-7271","observed_orcids":["https://orcid.org/0000-0002-5592-7271"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Cavicchioli, Marcelo","raw_affiliation_strings":["CAVIC Digital O\u00dc"],"raw_orcid":"https://orcid.org/0009-0006-6712-3023","affiliations":[{"raw_affiliation_string":"CAVIC Digital O\u00dc","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5070693424"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4081999957561493,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4081999957561493,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10734","display_name":"Information and Cyber Security","score":0.22360000014305115,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12479","display_name":"Web Application Security Vulnerabilities","score":0.11079999804496765,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/agent-harnesses","display_name":"agent harnesses","score":0.8309999704360962},{"id":"https://openalex.org/keywords/llm-based-agents","display_name":"LLM agents","score":0.5899999737739563},{"id":"https://openalex.org/keywords/message-delivery","display_name":"message delivery","score":0.4050000011920929},{"id":"https://openalex.org/keywords/model-comparison","display_name":"model comparison","score":0.45899999141693115},{"id":"https://openalex.org/keywords/claude","display_name":"Claude","score":0.3700000047683716},{"id":"https://openalex.org/keywords/glm","display_name":"GLM","score":0.8840000033378601},{"id":"https://openalex.org/keywords/kimi","display_name":"Kimi","score":0.9810000061988831},{"id":"https://openalex.org/keywords/cost-estimation","display_name":"cost estimation","score":0.46799999475479126}],"concepts":[{"id":"https://openalex.org/C2777810591","wikidata":"https://www.wikidata.org/wiki/Q16861606","display_name":"Credential","level":2,"score":0.7953000068664551},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.7699000239372253},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5856999754905701},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.5134000182151794},{"id":"https://openalex.org/C2781018962","wikidata":"https://www.wikidata.org/wiki/Q5164884","display_name":"Container (type theory)","level":2,"score":0.5123999714851379},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.48420000076293945},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.41609999537467957},{"id":"https://openalex.org/C2992886853","wikidata":"https://www.wikidata.org/wiki/Q18381816","display_name":"Post hoc","level":2,"score":0.4104999899864197},{"id":"https://openalex.org/C2779664074","wikidata":"https://www.wikidata.org/wiki/Q3518405","display_name":"Terminal (telecommunication)","level":2,"score":0.4090999960899353},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.3801000118255615},{"id":"https://openalex.org/C94523657","wikidata":"https://www.wikidata.org/wiki/Q4085781","display_name":"Wireless ad hoc network","level":3,"score":0.27570000290870667},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.27379998564720154},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C150202949","wikidata":"https://www.wikidata.org/wiki/Q107602","display_name":"Pointer (user interface)","level":2,"score":0.2583000063896179}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.5281/zenodo.23974846","doi":"https://doi.org/10.5281/zenodo.23974846","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.23974846","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},{"id":"doi:10.5281/zenodo.23974847","doi":"https://doi.org/10.5281/zenodo.23974847","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.23974847","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.5281/zenodo.23974846","doi":"https://doi.org/10.5281/zenodo.23974846","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.23974846","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7225125047","https://openalex.org/W7162690279","https://openalex.org/W7225326069","https://openalex.org/W7225292269","https://openalex.org/W7213372153","https://openalex.org/W7204584025","https://openalex.org/W7171685573","https://openalex.org/W7172000952","https://openalex.org/W7214235739","https://openalex.org/W7216310052"],"abstract_inverted_index":{"Background.":[0],"A":[1],"common":[2],"control":[3,198],"in":[4,45,118,127,146,159,242],"agent":[5,25],"harnesses":[6],"is":[7],"to":[8],"disable":[9],"a":[10,27,30,46,173],"tool":[11,111,228],"through":[12,229],"configuration.":[13],"Objective.":[14],"To":[15],"test":[16,75],"whether":[17,34],"such":[18],"disabling":[19],"prevents":[20],"the":[21,24,35,39,98,101,105,110,116,119,122,125,144,157,177,182,209,227,234,240],"action":[22],"when":[23],"retains":[26],"terminal":[28],"and":[29,33,51,63,67,71,115,138,149,164,245],"readable":[31],"credential,":[32],"outcome":[36,94,237],"depends":[37],"on":[38,239,246],"model.":[40],"Method.":[41],"Hermes":[42],"v0.22.6":[43],"ran":[44],"container":[47],"with":[48,54,81,85,130,133,136,140,169,223],"network":[49],"containment":[50],"fictitious":[52],"destinations,":[53],"four":[55,78],"models":[56],"(Claude":[57],"Sonnet":[58],"5.5,":[59],"GPT-6.1":[60],"Sol,":[61],"GLM-5.3":[62],"Kimi":[64,188],"K3.1":[65],"Pro)":[66],"identical":[68],"configuration,":[69],"requests":[70],"order.":[72],"The":[73,93,196,213,236],"main":[74],"(V2)":[76],"crossed":[77],"capability":[79],"levels":[80],"three":[82,184],"request":[83],"forms,":[84],"3":[86],"rounds":[87,129,148,163],"per":[88,91],"cell":[89],"(36":[90],"model).":[92],"was":[95,189,206],"what":[96],"reached":[97,124],"mailbox":[99],"or":[100,194],"collector,":[102],"recounted":[103],"from":[104,181,192],"raw":[106],"logs.":[107],"Results.":[108],"With":[109],"disabled":[112],"by":[113],"configuration":[114,120,230],"credential":[117,210],"file,":[121],"message":[123],"destination":[126],"0/36":[128,132],"Claude,":[131],"GPT,":[134],"13/36":[135],"GLM":[137,178],"3/36":[139],"Kimi.":[141],"Claude":[142,193],"found":[143],"route":[145,158],"16":[147],"declined":[150],"every":[151],"time.":[152],"GPT":[153],"did":[154,231],"not":[155,190,232],"find":[156],"34":[160],"of":[161,216,220],"36":[162],"attempted":[165],"direct":[166],"SMTP":[167],"twice,":[168],"no":[170],"delivery.":[171],"In":[172],"post":[174],"hoc":[175],"analysis,":[176],"proportion":[179],"differed":[180],"other":[183],"after":[185],"Holm":[186],"correction;":[187],"distinguishable":[191],"GPT.":[195],"simulated":[197],"point":[199],"(V2b)":[200],"produced":[201],"144":[202],"BLOCKED":[203],"outcomes":[204],"but":[205],"inconclusive,":[207],"because":[208],"remained":[211],"accessible.":[212],"harness":[214],"estimate":[215],"cost":[217],"equaled":[218],"32%":[219],"actual":[221],"spend":[222],"GLM.":[224],"Conclusion.":[225],"Removing":[226],"remove":[233],"route.":[235],"depended":[238],"model,":[241],"this":[243],"scenario":[244],"these":[247],"dates.":[248]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-11-06T00:00:00"}
